RE: [LC Comment ONT]

thank you for your responses

If you go with the current approach you should clearly indicate that the mappings you are covering on your document are only informative. One suggestion I would be in favour of would be to create one Recommendation with the Ontology (the normative staff) and one W3C Note with the mappings identified. That will make things clearer while at the same time will allow to evolve the Rec and the mappings independently. Also will make the Rec thinner and easier to be read.

best regards

De: Veronique Malaise [mailto:vmalaise@few.vu.nl]
Enviado el: lunes, 14 de junio de 2010 14:52
Para: Daniel Park
CC: JOSE MANUEL CANTERA FONSECA
Asunto: Re: [LC Comment ONT]


On Jun 14, 2010, at 1:45 PM, Daniel Park wrote:


Obviously, we are targeting for W3C Recommendation.

It seems a trade-off issue. Light version can be deployed quickly even if its mapping does not cover all of properties. That's why it's named as (Media Ontology 1.0). 2.0 and 3.0 will be able to expand mappings as wide as your expectation.

I think that the comment was about the fact that the mappings are recommended to be properties from the SKOS vocabulary, which is less formal than properties defined in OWL, hence the lightweight. Several projects in the Cultural Heritage world have opted for the less formal SKOS vocabulary, which is a W3C recommendation itself, several other projects, like the New York Times Topics mappings to DBpedia and GeoNames initiative, have discussed at length which option to choose from. I personaly am in favor of not over-committing the Media Ontology with strong mappings, particularly because of the difference of semantics (and syntax) between the elements of the different vocabularies/schemas we are seeking a mapping for. But I would of course be very interested to have Mr Cantera's opinion about the troubles I might not have foreseen? The Media Ontology describes the types of mappings that can be done, but we do only implementation examples, the text and mapping tables can be used to create a more constrained mapping if a user needs one. Maybe there was a misunderstanding on this level?

Best regards
Véronique




Daniel

--
Soohong Daniel Park
http://www.soohongp.com


On Mon, Jun 14, 2010 at 7:32 PM, JOSE MANUEL CANTERA FONSECA <jmcf@tid.es<mailto:jmcf@tid.es>> wrote:
Is this specification going to be a Rec or a Note?

because if this is going to be a Rec, you might be in trouble with your lightweight approach to mappings

best